Marko Krema

Text mining for product attribute extraction

We describe our work on extracting attribute and value pairs from textual product descriptions. The goal is to augment databases of products by representing each product as a set of attribute-value pairs. Such a representation is beneficial for tasks where treating the product as a set of attribute-value pairs is more useful than as an atomic entity. Examples of such applications include demand forecasting, assortment optimization, product recommendations, and assortment comparison across retailers and manufacturers. We deal with both implicit and explicit attributes and formulate both kinds of extractions as classification problems. Using single-view and multi-view semi-supervised learning algorithms, we are able to exploit large amounts of unlabeled data present in this domain while reducing the need for initial labeled data that is expensive to obtain. We present promising results on apparel and sporting goods products and show that our system can accurately extract attribute-value pairs from product descriptions. We describe a variety of application that are built on top of the results obtained by the attribute extraction system.

Predicting customer shopping lists from point-of-sale purchase data

This paper describes a prototype that predicts the shopping lists for customers in a retail store. The shopping list prediction is one aspect of a larger system we have developed for retailers to provide individual and personalized interactions with customers as they navigate through the retail store. Instead of using traditional personalization approaches, such as clustering or segmentation, we learn separate classifiers for each customer from historical transactional data. This allows us to make very fine-grained and accurate predictions about what items a particular individual customer will buy on a given shopping trip.We formally frame the shopping list prediction as a classification problem, describe the algorithms and methodology behind our system, its impact on the business case in which we frame it, and explore some of the properties of the data source that make it an interesting testbed for KDD algorithms. Our results show that we can predict a shoppers shopping list with high levels of accuracy, precision, and recall. We believe that this work impacts both the data mining and the retail business community. The formulation of shopping list prediction as a machine learning problem results in algorithms that should be useful beyond retail shopping list prediction. For retailers, the result is not only a practical system that increases revenues by up to 11%, but also enhances customer experience and loyalty by giving them the tools to individually interact with customers and anticipate their needs.

Building intelligent shopping assistants using individual consumer models

This paper describes an Intelligent Shopping Assistant designed for a shopping cart mounted tablet PC that enables individual interactions with customers. We use machine learning algorithms to predict a shopping list for the customers current trip and present this list on the device. As they navigate through the store, personalized promotions are presented using consumer models derived from loyalty card data for each inidvidual. In order for shopping assistant devices to be effective, we believe that they have to be powered by algorithms that are tuned for individual customers and can make accurate predictions about an individuals actions. We formally frame the shopping list prediction as a classification problem, describe the algorithms and methodology behind our system, and show that shopping list prediction can be done with high levels of accuracy, precision, and recall. Beyond the prediction of shopping lists we briefly introduce other aspects of the shopping assistant project, such as the use of consumer models to select appropriate promotional tactics, and the development of promotion planning simulation tools to enable retailers to plan personalized promotions delivered through such a shopping assistant.

Extracting and Using Attribute-Value Pairs from Product Descriptions on the Web

We describe an approach to extract attribute-value pairs from product descriptions in order to augment product databases by representing each product as a set of attribute-value pairs. Such a representation is useful for a variety of tasks where treating a product as a set of attribute-value pairs is more useful than as an atomic entity. We formulate the extraction task as a classification problem and use Naive Bayes combined with a multi-view semi-supervised algorithm (co-EM). The extraction system requires very little initial user supervision: using unlabeled data, we automatically extract an initial seed list that serves as training data for the semi-supervised classification algorithm. The extracted attributes and values are then linked to form pairs using dependency information and co-location scores. We present promising results on product descriptions in two categories of sporting goods products. The extracted attribute-value pairs can be useful in a variety of applications, including product recommendations, product comparisons, and demand forecasting. In this paper, we describe one practical application of the extracted attribute-value pairs: a prototype of an Assortment Comparison Tool that allows retailers to compare their product assortments to those of their competitors. As the comparison is based on attributes and values, we can draw meaningful conclusions at a very fine-grained level. We present the details and research issues of such a tool, as well as the current state of our prototype.